Conventionally, an image sensor has a plurality of reading modes (such as an addition reading mode and an HDR reading mode), and in some image sensors, the modes can be switched while capturing an image (for example, see Patent Document 1).
Meanwhile, in a case where a reading mode is changed in an image sensor, a shutter speed (exposure time) and time for reading pixel data (hereinafter also referred to as frame) are also changed. Consequently, there occurs, at timing of changing the reading mode, a corrupt frame which cannot be used as an image since exposure time for respective pixels is not uniform.
FIG. 1 illustrates a conventional method for coping with corrupt frames.
In other words, as illustrated in A in the figure, a first coping method is a method in which a corrupt frame, which may occur when a reading mode A (MODE-A) is changed to a reading mode B (MODE-B), is not output to a succeeding stage. According to the first coping method, it is possible to always output to a succeeding stage only a frame which can be used as an image. However, in the first coping method, output of frames is interrupted by one frame, and there is a demand to continuously output frames as long as it is possible. Accordingly, there is a second method of interest to meet the demand.
As illustrated in B in the figure, the second coping method is a method in which a frame synchronization signal (Vsync) is sent again when the reading mode A (MODE-A) is changed to the reading mode B (MODE-B).